Archive · 14 Nov 2025
Microsoft's Project Gecko Aims for Inclusive AI Development
Imagine computers that can talk and understand people, like a super-smart helper.
Right now, these computer helpers, called AI, are mostly trained using information from only a few places and in a few languages.
This means they don't work very well for many people around the world who speak different languages or have different cultures.
Microsoft has started a new project called 'Project Gecko' to fix this.
They want to build AI helpers that can understand and use local languages and are sensitive to different cultures.
They are also making these AI helpers smaller and more efficient so they can work on cheaper phones and computers.
This way, more people, like farmers in India and Kenya, can use them to get helpful information in their own language, helping them with their farms, and later, maybe even with school or health.
Microsoft has launched Project Gecko to develop cost-effective and tailorable AI systems for underrepresented communities.
The initiative focuses on creating AI that utilizes local languages, culturally sensitive content, and multimodal engagement (text, voice, video).
Project Gecko researchers are working with Small Language Models (SLMs) to ensure AI can run on low-capacity devices.
The project involves collaboration between Microsoft Research teams in Africa, India, and the US, along with Digital Green.
Initial applications are being developed for small farms in India and Kenya, with plans to expand into healthcare, education, and retail.
- Who
- Microsoft
- What
- Announced Project Gecko, an initiative to create inclusive and cost-effective AI systems using local languages and culturally sensitive content.
- When
- Not specified, but announced recently.
- Where
- Involving researchers from Microsoft Research Africa (Nairobi), Microsoft Research India, Microsoft Research Accelerator (USA), and Digital Green, with initial focus on India and Kenya.
- Why
- To address the poor performance of current AI in many languages and its failure to reflect diverse social and cultural realities, aiming to provide more innovative and useful AI solutions for underrepresented populations.
Addressing AI Inequality
Challenges in AI Development
Representation in AI
Addressing AI Inequality
Project Gecko aims to create AI systems that reflect the knowledge, languages, and cultures of the global majority, leading to more innovative and useful solutions for a wider range of people.
Challenges in AI Development
Current AI systems often perform poorly in many languages and do not accurately represent the social and cultural realities of all populations due to a lack of diverse training data.
Technological Accessibility
Addressing AI Inequality
Focuses on developing Small Language Models (SLMs) and cost-effective AI systems that can run on low-cost connected devices with limited computing capacity, common in the Global South.
Challenges in AI Development
Many low-cost connected devices in the Global South lack the necessary computing power to effectively run large, complex AI models.
- Project Name
- Project Gecko
- Initiative By
- Microsoft
- Key Focus
- Creating cost-effective, tailorable AI systems for underrepresented communities
- Developed Technology
- MMCTAgent (Multimodal Critical Thinking Agent Framework)
- AI Models Used
- Small Language Models (SLMs) and experimental frontier models
- Initial Target Areas
- Small farms in India and Kenya
- Future Target Areas
- Healthcare, education, and retail